• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.11
• /
• pp.4084-4104
• /
• 2021

Visual place recognition (VPR) is a fundamental yet challenging task of mobile robot navigation and localization. The existing VPR methods are usually based on some pairwise similarity of image descriptors, so they are sensitive to visual appearance change and also computationally expensive. This paper proposes a simple yet effective four-step method that achieves adaptive convolutional neural network (CNN) landmark matching for VPR. First, based on the features extracted from existing CNN models, the regions with higher significance scores are selected as landmarks. Then, according to the coordinate positions of potential landmarks, landmark matching is improved by removing mismatched landmark pairs. Finally, considering the significance scores obtained in the first step, robust image retrieval is performed based on adaptive landmark matching, and it gives more weight to the landmark matching pairs with higher significance scores. To verify the efficiency and robustness of the proposed method, evaluations are conducted on standard benchmark datasets. The experimental results indicate that the proposed method reduces the feature representation space of place images by more than 75% with negligible loss in recognition precision. Also, it achieves a fast matching speed in similarity calculation, satisfying the real-time requirement.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.41-51
• /
• 2022

Cables are critical components of cable-stayed bridges. A structural health monitoring system provides real-time cable tension recording for cable health monitoring. However, the measurement data involve multiple sources of variability, i.e., varying environmental and operational factors, which increase the complexity of cable condition monitoring. In this study, a one-class classification method is developed for cable condition assessment using Bayesian factor analysis (FA). The single-peaked vehicle-induced cable tension is assumed to be relevant to vehicle positions and weights. The Bayesian FA is adopted to establish the correlation model between cable tensions and vehicles. Vehicle weights are assumed to be latent variables and the influences of different transverse positions are quantified by coefficient parameters. The Bayesian theorem is employed to estimate the parameters and variables automatically, and the damage index is defined on the basis of the well-trained model. The proposed method is applied to one cable-stayed bridge for cable damage detection. Significant deviations of the damage indices of Cable SJS11 were observed, indicating a damaged condition in 2011. This study develops a novel method to evaluate the health condition of individual cable using the FA in the Bayesian framework. Only vehicle-induced cable tensions are used and there is no need to monitor the vehicles. The entire process, including the data pre-processing, model training and damage index calculation of one cable, takes only 35 s, which is highly efficient.

• Journal of Software Assessment and Valuation
• /
• v.16 no.2
• /
• pp.127-133
• /
• 2020

As the usages of artificial intelligence is increased, the demand to algorithms for small portable devices increases. Also as the embedded system becomes high-performance, it is possible to implement algorithms for high-speed computation and machine learning as well as operating systems. As the machine learning algorithms process repetitive calculations, it depend on the cloud environment by network connection. For an stand alone system, low power consumption and fast execution by optimized algorithm are required. In this study, for the purpose of smart control, an energy measurement sensor is connected to an embedded system, and a real-time monitoring system is implemented to store measurement information as a database. Continuously measured and stored data is applied to a learning algorithm, which can be utilized for optimal power control, and a system interfacing various sensors required for energy measurement was constructed.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.12
• /
• pp.3836-3854
• /
• 2022

The student group division is benefit for universities to do the student management based on the group profile. With the widespread use of student smart cards on campus, especially where students living in campus residence halls, students' daily activities on campus are recorded with information such as smart card swiping time and location. Therefore, it is feasible to depict the students with the daily activity data and accordingly group students based on objective measuring from their campus behavior with some regular student attributions collected in the management system. However, it is challenge in feature representation due to diverse forms of the student data. To effectively and comprehensively represent students' behaviors for further student group division, we proposed to adopt activity data from student smart cards and student attributes as input data with taking account of activity and attribution relationship types from different perspective. Specially, we propose a novel student group division method based on a multi-view student attribute heterogeneous information network (MSA-HIN). The network nodes in our proposed MSA-HIN represent students with their multi-dimensional attribute information. Meanwhile, the edges are constructed to characterize student different relationships, such as co-major, co-occurrence, and co-borrowing books. Based on the MSA-HIN, embedded representations of students are learned and a deep graph cluster algorithm is applied to divide students into groups. Comparative experiments have been done on a real-life campus dataset collected from a university. The experimental results demonstrate that our method can effectively reveal the variability of student attributes and relationships and accordingly achieves the best clustering results for group division.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.3
• /
• pp.63-68
• /
• 2022

There are 395,789 people with hearing impairment in Korea, according to the 2021 Disability Statistics Annual Report by the Korea Institute for the Development of Disabled Persons. These people are experiencing a lot of inconvenience through hearing impairment, and many studies related to recognition and translation of Korean sign language are being conducted to solve this problem. In sign language recognition and translation research, collecting sign language data has many difficulties because few people use sign language professionally. In addition, most of the existed data is sign language data taken from the front of the speaker. To solve this problem, in this paper, we designed and developed a storage system that can collect sign language data based on multi-view points in real-time, rather than a single point, and store and manage it with high usability.

• Journal of the Society of Naval Architects of Korea
• /
• v.58 no.6
• /
• pp.366-374
• /
• 2021

This paper presents the analysis of the speed-power performance in the real sea using a large container vessel data provided as a test bed from a shipping company. To perform a speed trial of the vessel during a commercial voyage, the on-board measuring device and various operation data acquisition systems were mounted on the vessel for long-term performance monitoring and the voyage operated under the container loading condition close to the design draft was adopted. The content of this paper consists of Part I and Part II. Part I, such as this paper, contains the speed trial method and analysis results of the operating vessel. Part II contains the analysis of the speed-power characteristics change over time and before and after hull cleaning using operation data measured from the voyage operated under a condition similar to the speed trial.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.2
• /
• pp.51-58
• /
• 2022

Blockchain technology is a system in which data from users participating in blockchain networks is distributed and stored. Bitcoin and Ethereum are attracting global attention, and the utilization of blockchain is expected to be endless. However, the need for blockchain data privacy protection is emerging in various financial, medical, and real estate sectors that process personal information due to the transparency of disclosing all data in the blockchain to network participants. Although studies using smart contracts, homomorphic encryption, and cryptographic key methods have been mainly conducted to protect existing blockchain data privacy, this paper proposes data privacy using matrix character relocation techniques differentiated from existing papers. The approach proposed in this paper consists largely of two methods: how to relocate the original data to matrix characters, how to return the deployed data to the original. Through qualitative experiments, we evaluate the safety of the approach proposed in this paper, and demonstrate that matrix character relocation will be sufficiently applicable in private blockchain environments by measuring the time it takes to revert applied data to original data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.12
• /
• pp.4345-4363
• /
• 2021

Deep Learning as a Service (DLaaS), utilizing the cloud-based deep neural network models to provide customer prediction services, has been widely deployed on mobile cloud computing (MCC). Such services raise privacy concerns since customers need to send private data to untrusted service providers. In this paper, we devote ourselves to building an efficient protocol to classify users' images using the convolutional neural network (CNN) model trained and held by the server, while keeping both parties' data secure. Most previous solutions commonly employ homomorphic encryption schemes based on Ring Learning with Errors (RLWE) hardness or two-party secure computation protocols to achieve it. However, they have limitations on large communication overheads and costs in MCC. To address this issue, we present LeHE4SCNN, a scalable privacy-preserving and communication-efficient framework for CNN-based DLaaS. Firstly, we design a novel low-expansion rate homomorphic encryption scheme with packing and unpacking methods (LeHE). It supports fast homomorphic operations such as vector-matrix multiplication and addition. Then we propose a secure prediction framework for CNN. It employs the LeHE scheme to compute linear layers while exploiting the data shuffling technique to perform non-linear operations. Finally, we implement and evaluate LeHE4SCNN with various CNN models on a real-world dataset. Experimental results demonstrate the effectiveness and superiority of the LeHE4SCNN framework in terms of response time, usage cost, and communication overhead compared to the state-of-the-art methods in the mobile cloud computing environment.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.1
• /
• pp.205-212
• /
• 2023

In the modern society, the extreme weather caused by climate change has brought about exceptional damage in succession over the world due to the use of fossil fuels, and infectious diseases such as COVID-19 worsen the quality of human life. It is urgently necessary to reduce green-house gas and use new renewable energy. The global environmental pollution should be decreased by reducing the use of fossil fuels and using new renewable energy. This paper suggests a system which can function for the environment of four seasons, safety and communication, through the photovoltaic power-based intelligent CCTV, internet and WiFi, and cooling and heating systems, and can optimally manage power, through the real-time monitoring of the production and the consumption of the photovoltaic power. It suggests a hybrid generation system supporting diesel generation without discontinuation in the case of emergency such as system power outage caused by cold waves, typhoons and natural disasters in which the photovoltaic power generating system cannot be used.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.5
• /
• pp.422-428
• /
• 2023

Ultrafast ultrasound imaging has been applied to various imaging approaches, including shear wave elastography, ultrafast Doppler, and super-resolution imaging. However, these methods are still challenging in real-time implementation for three Dimension (3D) or portable applications because of their massive data rate required. In this paper, we proposed an adaptive quantization method that effectively reduces the data rate of large Radio Frequency (RF) data. In soft tissue, ultrasound backscatter signals require a high dynamic range, and thus typical quantization used in the current systems uses the quantization level of 10 bits to 14 bits. To alleviate the quantization level to expand the application of ultrafast ultrasound imaging, this study proposed a depth-sectional quantization approach that reduces the quantization errors. For quantitative evaluation, Field II simulations, phantom experiments, and in vivo imaging were conducted and CNR, spatial resolution, and SSIM values were compared with the proposed method and fixed quantization method. We demonstrated that our proposed method is capable of effectively reducing the quantization level down to 3-bit while minimizing the image quality degradation.